Neurodevelopment at 5 years of age for preterm-born children according to mode of conception: a cohort study

BACKGROUND: Preterm delivery is a risk factor for suboptimal neu-rodevelopment. Pregnancies conceived after medically assisted repro-duction—which includes in vitro fertilization, with or without intracytoplasmic insemination, and induction of ovulation followed by intrauterine insemination or timed intercourse—have a higher risk of pre-term delivery. Few studies have evaluated the outcome at > 2 years of age of such preterm-born children. OBJECTIVE: To evaluate neurodevelopmental outcome at 5 1 / 2 years of age of children born preterm according to the mode of conception (spontaneous vs


Introduction
Infertility, defined as inability to conceive after 1 year of regular, unprotected sexual intercourse, 1 affects approximately 15% of couples. 2 This has many consequences, including sexual dysfunction, social stigmatization, and relationship breakdown.3e8 Since the birth of the first child conceived through in vitro fertilization (IVF) in 1978, the use of assisted reproductive technologies (ARTs) has increased substantially, such that 3% to 4% of births are now from pregnancies conceived through ART. 9,10owever, ART includes only the in vitro handling of oocytes and sperm, or of embryos, thus including IVF and IVF with intracytoplasmic sperm injection (ICSI) but not intrauterine insemination (IUI) following induction of ovulation (IO) or timed intercourse (TIC) following IO. 1 These techniques, which fall under the broader term of "medically-assisted reproduction" (MAR), 1 also expose women and fetuses to exogenous hormones.
A major concern for women undergoing MAR and their partners is longterm neurodevelopmental outcome of the offspring.Problems potentially arise because genes subject to parental imprinting may be affected by epigenetic modifications relating to MAR (hormonal IO, manipulation of male gametes, IVF, or embryo transfer), thus negatively affecting the offspring. 11MAR is also associated with both an increased risk of preterm birth (at <37 weeks of gestational age [GA]), including very preterm birth (at <32 weeks' GA), 12 and with multiple pregnancy (particularly following multifollicular stimulation or multiple embryo transfers), which is itself associated with preterm birth. 10reterm birth is in turn associated with a risk of poorer neurodevelopmental outcomes.13e15 To date, however, data concerning the neurodevelopment of children born following MAR have been inconsistent, with studies finding outcomes in children conceived following MAR compared with children conceived naturally to be poorer, 16 better, 17,18 or the same. 12,19Such discrepancies arise because of differences in neurodevelopmental domains studied and age at follow-up, and methodological differences between the studies 20 ; the true impact of MAR therefore remains unclear. 12hen looking more specifically at neurodevelopment in children born preterm following MAR, few data are available.Two retrospective populationbased studies of births before 29 weeks' GA assessed at 18 to 24 months found conflicting results, 18,21 whereas a singlecenter study of births before 34 weeks found a reduced probability of poor neurodevelopment at 2 years of age. 22nly 1 study has examined outcomes after at least 5 years of age-and only in relation to cerebral palsy-and found no differences by mode of conception. 23owever, neurodevelopment is dynamic and evolves over time: motor deficits become apparent first, with cognitive deficits appearing later.By age 5, more subtle defects are detectable, thus multiple dimensions of development should be studied.We sought to evaluate the impact of mode of conception on neurodevelopment at 5 1 / 2 years of age in children born at <35 weeks' GA.Our primary objective was to assess whether any effect on neurodevelopment was evident using all MAR techniques combined.Different techniques may also have different effects: IVF or IVF-ICSI techniques, such as multifollicular stimulation, gamete manipulation, embryo culture, and embryo transfer, might cause epigenetic disturbances; these might also be observed following hormonal stimulation (IO with TIC, or IUI). 11We therefore studied subgroups of children born following the use of these techniques in comparison with those born after spontaneous conception.We hypothesized that there would be no differences once social factors were accounted for.

Setting and data collection
The French prospective, national cohort study "EPIPAGE-2" collected information about all births at <35 weeks' GA in 546 maternity hospitals in France in 2011. 24,25Children born at <27 weeks' GA were recruited over 8 months (equivalent to 35 weeks), those born between 27 and 31 completed weeks of GA over 6 months (equivalent to 26 weeks), and those between 32-and 34weeks' GA over 5 weeks.At birth, maternal, obstetrical, and neonatal data were obtained from medical records, and during the child's hospital stay, mothers were interviewed to obtain information on their social characteristics and pregnancy.Surviving children were seen by trained investigators at 5 1 / 2 years of age: this included a medical examination and neuropsychological assessment, and parents completed a questionnaire.

Population
Only children born between 24-and 34weeks' GA were included because the 1 child born at <24 weeks who survived was lost to follow-up at 5 1 / 2 years.We excluded children for whom mode of conception or, if born following MAR, type of infertility treatment were unknown.

Exposure
Birth following MAR was compared with that following spontaneous conception.Information about MAR was collected at birth from medical notes and postnatal interview; the use of hormonal stimulation (IO with TIC, or IUI) or IVF (alone or with ICSI) were accepted as evidence of an MARconceived pregnancy.The subgroups of IO with TIC or IUI, and IVF or IVF-ICSI were also examined separately.

Main outcomes
We studied cerebral palsy, sensory (hearing and vision) and cognitive impairments, and developmental coordination disorders.Cerebral palsy was diagnosed clinically using the Surveillance of Cerebral Palsy in Europe network criteria and classified according to the Gross Motor Function Classification System (GMFCS).Visual impairment was defined as binocular visual acuity <3.2/10, and hearing impairment was defined as uni-or bilateral hearing loss >40 dB not corrected or only partially corrected with hearing aids.Cognitive ability was measured using the full-scale intelligence quotient (FSIQ) from the Wechsler Preschool and Primary Scale of Intelligence-Fourth Edition (WPPSI-IV, French version) 26 ; this composite score is obtained from 5 domains: verbal comprehension, visuospatial indices, fluid reasoning, working memory, and processing speed.We studied mean intelligence quotient (IQ) and proportions of children with scores both 1 and 2 standard deviations (SDs) below the mean of a reference group of AJOG at a Glance Why was this study conducted?Some studies have suggested that children born preterm following medically assisted reproduction (MAR) are at additional risk for neurodevelopmental impairments.These have mostly focused on outcomes at up to 2 years of age, but evidence at older ages is lacking.

Key findings
In this prospective cohort, there were no differences in neurodevelopmental outcomes-including cognition, cerebral palsy, combined neurosensory impairment, and developmental coordination disorders-at 5 1 / 2 years of age between children born preterm after MAR and those conceived naturally when results were fully adjusted for sociodemographic factors.

What does this add to what is known?
MAR is not associated with additional long-term neurodevelopmental impairments at up to 5 1 / 2 years of age for children born preterm.
Original Research GYNECOLOGY ajog.org1.e2 American Journal of Obstetrics & Gynecology MONTH 2022 children born at term (37e41 weeks' GA).The reference group, weighted to be representative of the French population, was selected from a contemporaneous birth cohort, with children undergoing the same follow-up as those in EPIPAGE-2. 27Moderate to severe impairment was defined as having at least 1 sensory impairment, cerebral palsy GMFCS level 2 or higher, or FSIQ >2 SDs below the mean of the reference group.Finally, developmental coordination disorders were assessed in children without moderate or severe impairment using the Movement Assessment Battery for Children, Second Edition.We studied both total scores and using a cutoff score below the fifth percentile (relative to the reference group).

Other studied factors
Maternal characteristics available from delivery were: maternal age (years), level of education (less than high school, high school, 1e2 years or >2 years of graduate study), currently employed (yes/ no), birth country (France or elsewhere), cohabitation status, smoking during pregnancy (yes/no), and household socioeconomic status defined according to the highest status of the mother and partner, or mother only if she lived alone (executive, intermediate, administration, service and trade, manual worker, and unemployed).Obstetrical variables were: parity (nulliparous or not), singleton or multiple pregnancy, induced or spontaneous labor, delivery mode (vaginal or cesarean), receipt of antenatal steroids and tocolysis, and neonatal unit level at delivery hospital.Neonatal characteristics were: GA at delivery (completed weeks), sex, small for GA (using French reference curves 28 ), and severe malformations (yes/no).

Statistical analysis
Population characteristics were described using means and SDs for continuous variables, with groups compared using Student or Wilcoxon tests.For categorical variables, we described proportions and used chi square or Fisher exact tests.Mortality rates between birth and follow-up were assessed to determine whether there were differences in survival between children conceived following MAR and those conceived spontaneously.Regression analyses were performed among survivors aged 5 1 / 2 years.For binary outcomes, odds ratios were estimated with logistic regression, and for continuous outcomes, linear regression was used; both used generalized estimating equations (GEE) to account for nonindependence of multiple children born to the same mother.We first estimated crude associations, then added GA at birth and antenatal steroids to explore any potential mediating impact from these factors, and finally, in the fully adjusted model, included sociodemographic variables that were considered a priori to be potential confounders.These were: maternal age, parity, birth country, level of education, employment, smoking during pregnancy, cohabitation, and socioeconomic status.A P value <.05 was considered statistically significant; results are presented with 95% confidence intervals (CIs).
Data were weighted according to GA group by a factor of 1.34 (35/26) for children born at 27 to 31 weeks and by a factor of 7 (35/5) for those born at 32 to 34 weeks to account for the differing recruitment periods, and multiple imputation was performed using chained equations to account for missing outcome data; imputation models included variables potentially predicting nonresponse or the outcome (Supplemental Table 1).Estimates were combined using Rubin's rules. 29All investigations were conducted using R, version 4.1.1(R Foundation for Statistical Computing, Vienna, Austria). 30EE models were performed using the R package "geepack" 31 and multiple imputation with the package "mice." 32

Sensitivity analyses
We did not adjust for multiple pregnancy status because this potentially lies on the causal pathway between MAR and later outcome.Instead, we repeated analyses using singleton births only.We also performed sensitivity analyses of all models on the population of children with complete data.

Study population
At 24 to 34 completed weeks' GA, 5022 children were born alive; mode of conception, including type of infertility treatment, was known for 4907.Of these, 558 children died before reaching 5 1 / 2 years of age (4.6% in the MAR and 6.0% in the non-MAR group, P¼.025).Among the 4349 survivors, complete information was available for 641 of 814 children born following MAR and 2390 of 3535 children born following spontaneous pregnancy.More children were lost to follow-up from spontaneous and singleton pregnancies and from families of lower socioeconomic status, and their mothers were more often younger, multiparous, single, smokers, and born outside of France (Figure; Supplemental Tables 2 and 3).

Baseline characteristics
MAR was used by 532 mothers; at delivery, these mothers were older, more highly educated, more often employed, of higher socioprofessional category, and less frequently smokers than women with spontaneous pregnancies (Table 1).Their children were more often from multiple pregnancies, small for GA, exposed to antenatal steroids, and born following spontaneous labor, with noncephalic presentation, and in hospitals with level 3 neonatal intensive care units.There were no differences in terms of GA, sex, presence of severe malformations, or mode of delivery (Table 2).
Outcomes at 5 1 / 2 years of age At 5 1 / 2 years of age, before adjustment, there were better outcomes for cognition among children born following MAR than among those from spontaneous pregnancies (Table 3).These differences disappeared following adjustment with sociodemographic variables (Table 4).Similar results were observed for the composite measure of moderate to severe neurodevelopmental impairment.There were no differences in unadjusted or adjusted analyses for cerebral palsy, nor in proportions of sensory deficiencies, between children born following MAR and children from spontaneous pregnancy (Table 3), nor were any differences identified for developmental coordination disorders (Tables 3 and 4).

Subgroup analyses
For both the children born following IO or IUI and those born following IVF or IVF-ICSI, similar patterns to those observed in the main analysis were noted.In both groups, the mean IQ before adjustment was higher than that of children born from spontaneous pregnancies, and fewer children had an IQ <1 SD (Table 3); the proportion of children with an IQ <2 SDs was also lower in the IVF/IVF-ICSI group but not in the IO/IUI group.Again, after adjustment for sociodemographic factors, no differences persisted (Table 4).

Sensitivity analyses
Results for singleton analyses were consistent for all outcomes among the entire population (Supplemental Table 4).In subgroup analyses, the odds ratio for having an FSIQ <1 SD Original Research GYNECOLOGY ajog.org1.e4 American Journal of Obstetrics & Gynecology MONTH 2022 below the mean was lower among those born following IO or IUI than among those born after spontaneous conception, but all other results were nonsignificant after adjustment for sociodemographic factors and consistent with analyses of the full population.Results from the IVF/IVF-ICSI group were also consistent with the main results (Supplemental Table 4), with no differences seen in complete case analyses (Supplemental Tables 5 and 6).

Principal findings
In this prospective cohort study of preterm-born children followed-up at 5 1 / 2 years of age, we found no evidence of an association between mode of conception and moderate to severe neurodevelopmental impairment following adjustment for sociodemographic factors, nor with cerebral palsy, sensory impairments, cognition, or developmental coordination disorders.Results were the same regardless of whether outcomes were analyzed as continuous scores or in binary categories representing potentially serious impairment.Sensitivity analyses using singletons and complete cases were also consistent.

Results in the context of what is known
Previous studies have identified differences in neurodevelopment related to mode of conception among children born preterm, particularly for cerebral palsy.One study that found an increased risk of cerebral palsy following IVF included children born between 1982 and 1995; however, evaluation was done at 2 years of age and results were only adjusted for the child's sex, year of birth, and maternal age 16 ; moreover, IVF techniques have evolved since then. 12ncreased risk of cerebral palsy was identified for children born at <32 weeks' GA in a whole-population Australian study, but CIs were wide because few very preterm children were included. 33 different, prospective Australian cohort had similar findings, with an increased risk of moderate to severe neurodevelopmental impairment at 2 to 3 years of age for children born between  22 ; similar results were found in a retrospective study adjusting for maternal education.18 However, neither study accounted for missing data, and the results are thus difficult to interpret.Other studies were also restricted to complete-case analyses.16,33 Not only did we use multiple imputation and perform sensitivity analyses on complete cases, but we also adjusted for multiple social factors, and found that any initially perceived differences in outcome following MAR disappeared following this adjustment.

Clinical implications
The finding that there are no differences in neurodevelopmental outcomes according to mode of conception in this prospectively collected French national cohort of very and moderately preterm children is highly likely to apply also in other countries.External validity may be limited because French perinatal care was less active than that of some other countries for neonates born extremely preterm (at <27 weeks' GA), but it was not dissimilar to that of other European countries 34 ; furthermore, these children represented only approximately 5% of the births included in this study.Of greater concern might be that MAR techniques have changed: methods for both freezing embryos and the media in which they were subsequently cultured were different in 2011, and the transfer of several embryos was also more frequent and usually occurred at day 2 or 3. 10 However, evolution of practice has occurred internationally, not just within France, and longer-term follow-up necessarily requires that practices are   from the past, thus implying that they are likely to have evolved in the interim period.Consequently, this study should be reassuring for health professionals and parents or parents-to-be of children born preterm following MAR because it indicates that any developmental consequences arise from preterm birth rather than the mode of conception itself conveying an additional risk.

Research implications
Although EPIPAGE-2 contains a wealth of follow-up and social data, information was limited about the MAR techniques.We did not have details about which drugs were used and at what dose, or whether embryos were transferred fresh or frozen, nor could we identify children born from donated gametes.This leaves questions about the impact of more specific fertility treatments for future research.

Strengths and limitations
This study evaluated multiple dimensions of longer-term neurodevelopment among children born preterm according to the mode of conception.Using a large, prospectively collected national cohort with comprehensive data covering a range of medical and sociodemographic characteristics 24,25 allowed us to study several neurodevelopmental outcomes while taking into consideration important confounding factors with sufficient power to detect potential differences, particularly for the most frequent outcomes (cognitive impairment, developmental coordination disorder, and cerebral palsy).The quality of the used sociodemographic information is a further strength: most previous studies had only medical data with minimal additional information, and given that the social environment is a major predictor of child development, residual confounding may have been an issue.The main difficulty in prospective cohort studies is loss to follow-up. 35Data we had available covered pregnancy, the neonatal hospitalization, and subsequent course of the children, thus allowing us to use these data in imputation models, thereby increasing the likelihood of the "missing at random" assumption being met. 32This is important because missing data may have impact in ways that are difficult to determine. 35We were further reassured by the very similar results found in analyses using complete cases.We were also able to examine the implication of broader MAR techniques both together and separated into ART and non-ART techniques, although more detailed information about specific techniques was not available; previous studies in the preterm population have predominantly focused on children born either after IVF/IVF-ICSI 16,36 or after all types of MAR combined. 18,21,22Only 1 other study separated ARTand non-ART techniques, but it only examined the relationship with cerebral palsy and was restricted in the sociodemographic variables available for inclusion because the data were obtained from national registers. 23Our study was also limited by its restriction to preterm-born children; we are therefore only able to state that there was no increased risk of neurodevelopmental impairment according to mode of conception in this population, but it is also important to remember that there are increased risks of multiple pregnancy and preterm birth with the use of MAR. 10,12

Conclusion
In summary, we assessed neurodevelopmental outcomes at 5 1 / 2 years of age for children born preterm following MAR, and after adjusting for social characteristics, found no differences from children born following spontaneously conceived pregnancies.These are important insights for obstetricians, pediatricians, and other healthcare professionals working with women and their families.Our study provides further evidence for health professionals to reassure parents or parents-to-be when a child conceived from MAR is born preterm.n -As an oral presentation at the 8th congress of the European Academy of Paediatric Societies -EAPS 2020, held virtually, October 16e19, 2020.

Members of the EPIPAGE-2 Obstetric Writing Group
Data availability: The data are, in principle, accessible to all research teams, public, French, or foreign, subject to authorization by the cohort Data Access Committee.
The new law for modernization of the French Public Health System voted in 2016 now provides a legal framework for access to and reuse of already collected cohort data by complying with 'Reference Methodology MR-004.'Therefore, only nonnominative data defined as having a low reidentification risk are accessible.Moreover, general information on research activities in the institution must be provided to the persons concerned (posting on the premises, entry in the welcome booklet, etc.).
To this general information, individual patient information must be delivered for each project in which the patient is involved or for which the patient data will be treated.
As a consequence, each data access request must be submitted to the EPIPAGE-2 Data Access Committee (DAC) that evaluates the research projects on the basis of the following criteria: (1) methodological strengths and weaknesses (feasibility, choice of methods to achieve the objectives), (2) absence of overlap with other ongoing projects, in which case discussions with the different teams are organized, and (3) relevance of the requested data for the project and respect for confidentiality.

ajog.org 1 .
e12 American Journal of Obstetrics & Gynecology MONTH 2022 Comparison of children participating and nonparticipating in follow-up (N[4349 live children at 5.5 years)

TABLE 1
Characteristics of the 3667 mothers with children in the EPIPAGE-2 cohort surviving to 5 1 / 2 years of age according to mode of conception, after multiple imputation

TABLE 2
Pregnancy and childbirth outcomes for 4349 children from the EPIPAGE-2 cohort surviving at 5 1 / 2 years of age according to mode of conception, after multiple imputation Verhaeghe.Neurodevelopment at age 5 for preterm children born following medically assisted reproduction.Am J Obstet Gynecol 2022.
28ta are percentage (95% confidence interval) unless otherwise noted.Results are given after multiple imputation and are weighted to take into account the differences in survey design between gestational age groups; proportions are not exactly n/N because of the weighting.ICSI, intracytoplasmic sperm injection; IO, induction of ovulation; IUI, intrauterine insemination; IVF, in vitro fertilization; MAR, medically assisted reproduction.aMARcorrespondsto the whole range of MAR techniques, that is, IO, IUI, IVF, and IVF-ICSI; b Small-for-gestational-age was defined as birthweight <10th percentile for gestational age and sex on the basis of French intrauterine growth curves.281.e6American Journal of Obstetrics & Gynecology MONTH 2022

TABLE 3
Neurodevelopmental outcome measures for 4349 children from the EPIPAGE-2 cohort surviving at 5 1 / 2 years according to mode of conception, after multiple imputation Blindness or binocular corrected visual acuity <3.2/10; d Deafness, hearing loss >40 dB not corrected or partially corrected with hearing aid; e Full-scale intelligence quotient, measured by the Wechsler Preschool and Primary Scale of Intelligence-Fourth Edition 26 ; f Cutoff of the distribution related to a reference group born at term 27 ; g Severe or moderate cerebral palsy, severe or moderate sensory impairments, or FSIQ <2 SDs below the mean of a reference population; h Among children without cerebral palsy or severe or moderate sensory impairments, and with full-scale intelligence quotients !2 SDs below the mean of a reference population.Verhaeghe.Neurodevelopment at age 5 for preterm children born following medically assisted reproduction.Am J Obstet Gynecol 2022.
Data are percentage (95% confidence interval) unless otherwise noted.Results are given after multiple imputation and are weighted to take into account the differences in survey design between gestational age groups; proportions are not exactly n/N because of the weighting.FSIQ, full-scale intelligence quotient; ICSI, intracytoplasmic sperm injection; IO, induction of ovulation; IUI, intrauterine insemination; IVF, in vitro fertilization; MABC-2, Movement Assessment Battery for Children, Second Edition (Henderson, 2007); MAR, medically assisted reproduction; SD, standard deviation.aMARcorresponds to the whole range of MAR techniques, that is, IO, IUI, IVF, and IVF-ICSI; b Chi square test P value, estimated with the generalized estimating equations (GEE) approach to take into account correlation between twins or triplets, compared with spontaneous pregnancy; cGYNECOLOGY Original ResearchMONTH 2022 American Journal of Obstetrics & Gynecology 1.e7

TABLE 4
Association between mode of conception and neurodevelopmental outcome measures for 4349 children from the EPIPAGE-2 cohort surviving at 5 1 / 2 years-multivariate analysis after multiple imputation MAR corresponds to the whole range of MAR techniques, that is, IO, IUI, IVF, and IVF-ICSI; b The reported measures of association are odds ratios, except for FSIQ and total MABC-2 scores, where mean differences are reported.The generalized estimating equations approach was used to take into account correlation between twins or triplets; Verhaeghe.Neurodevelopment at age 5 for preterm children born following medically assisted reproduction.Am J Obstet Gynecol 2022.
CI, confidence interval; FSIQ, full-scale intelligence quotient; GA, gestational age; ICSI, intracytoplasmic sperm injection; IO, induction of ovulation; IUI, intrauterine insemination; IVF, in vitro fertilization; MABC-2, Movement Assessment Battery for Children, Second Edition (Henderson, 2007); MAR, medically assisted reproduction; OR, odds ratio; SD, standard deviation.a c Sociodemographic factors adjusted for are: maternal age, parity, education level, employment status, living with a partner, smoking during pregnancy, country of birth, and parents' socioeconomic status; d Full-scale intelligence quotient, measured by the Wechsler Preschool and Primary Scale of Intelligence-Fourth Edition 26 ; e Cutoff of the distribution related to a reference group born at term 27 ; f Severe or moderate cerebral palsy, severe or moderate sensory impairment, or FSIQ <2 SDs below the mean of a reference population; g Among children without cerebral palsy or severe or moderate sensory impairment, and with FSIQ !2 SDs below the mean of a reference population.1.e8 American Journal of Obstetrics & Gynecology MONTH 2022

TABLE 5
28mparison of children participating and nonparticipating in follow-up (N[4349 live children at 5.5 years) (continued)Data are presented as number of events/number in groups or percentages, unless otherwise indicated.For observed data, denominators vary according to the number of missing data for each variable.Results are weighted to consider the differences in survey design among GA groups.Proportions are not exactly number/total number because of weighting.GA, gestational age; ICSI, intracytoplasmic sperm injection; IO, induction of ovulation; IUI, intrauterine insemination; IVF, in vitro fertilization; MAR, medically assisted reproduction; SD, standard deviation; SGA, small for gestational age.MAR corresponds to the whole range of MAR techniques (ie, IO, IUI, IVF, and IVF-ICSI); b Defined as the highest occupational status among occupations of the mother and the father or mother only if living alone; c SGA was defined as a birthweight of <10th percentile for GA and sex based on French intrauterine growth curves.28Verhaeghe.Neurodevelopment at age 5 for preterm children born following medically-assisted reproduction.Am J Obstet Gynecol 2022.Survival at 5.5 years according to mode of conception (N[4907 live births included)Data are presented as number of events and percentages.Denominators vary according to the number of missing data for each variable.Results are weighted to consider the differences in survey design among gestational age groups.Proportions are not exactly number/total number because of weighting.ICSI, intracytoplasmic sperm injection; IO, induction of ovulation; IUI, intrauterine insemination; IVF, in vitro fertilization; MAR, medically assisted reproduction.MAR corresponds to the whole range of MAR techniques (ie, OI, IUI, IVF, and IVF-ICSI); b Chi-squared test P value, estimated with generalized estimating equations approach to consider the correlation between twins and triplets, compared to spontaneous pregnancy.Verhaeghe.Neurodevelopment at age 5 for preterm children born following medically-assisted reproduction.Am J Obstet Gynecol 2022.Association between mode of conception and neurodevelopmental outcome measures at 5.5 years, multivariate analysis after multiple imputation, among singleton pregnancies GA, gestational age; ICSI, intracytoplasmic sperm injection; IQ, intelligence quotient; IUI, intrauterine insemination; IVF, in vitro fertilization; MABC-2, Movement Assessment Battery for Children-Second Edition (Henderson, 2007); MAR, medically assisted reproduction; MD, mean difference; OI, ovulation induction; OR, odds ratio; SD, standard deviation.MAR corresponds to the whole range of MAR techniques (ie, OI, IUI, IVF, and IVF-ICSI); b The reported measures of association are OR, except for Full Scale IQ and total MABC-2 score where MDs are reported; c Sociodemographic factors adjusted for were maternal age, parity, education level, employment status, living with a partner, smoking during pregnancy, country of birth, and parents' socioeconomic status; d Full Scale IQ, measured using the Wechsler Preschool and Primary Scale of Intelligence-Fourth Edition 26 ; e Cutoff of the distribution related to a reference group born at term 27 ; f Severe or moderate cerebral palsy, severe or moderate sensory disabilities, or Full Scale IQ <2 SDs below the mean of a reference population; g Among children without cerebral palsy, without severe or moderate sensory disabilities, and with Full Scale IQ !2 SDs below the mean of a reference population.Verhaeghe.Neurodevelopment at age 5 for preterm children born following medically-assisted reproduction.Am J Obstet Gynecol 2022.Neurodevelopmental outcome measures at 5.5 years according to mode of conception (complete case analysis) Data are presented as number of events/number in groups or percentages, unless otherwise indicated.Denominators vary according to the number of missing data for each variable.Results are weighted to consider the differences in survey design among gestational age groups.Proportions are not exactly number/total number because of the weighting.Blindness or binocular corrected visual acuity of <3.2/10; d Deafness, hearing loss of >40 dB not corrected or partially corrected with hearing aid; e Full Scale IQ, measured by the Wechsler Preschool and Primary Scale of Intelligence-Fourth Edition 26 ; f Cutoff of the distribution related to a reference group born at term 27 ; g Severe or moderate cerebral palsy, severe or moderate sensory impairment, or Full Scale IQ quotient <2 SDs below the mean of a reference group; h Among children without cerebral palsy, without severe or moderate sensory impairment, and with Full Scale IQ !2 SDs below the mean of a reference group.Verhaeghe.Neurodevelopment at age 5 for preterm children born following medically-assisted reproduction.Am J Obstet Gynecol 2022.MONTH 2022 American Journal of Obstetrics & Gynecology 1.e19 Verhaeghe.Neurodevelopment at age 5 for preterm children born following medically-assisted reproduction.Am J Obstet Gynecol 2022.(continued)ajog.orgGYNECOLOGY Original Research MONTH 2022 American Journal of Obstetrics & Gynecology 1.e15 a a a 1.e18American Journal of Obstetrics & Gynecology MONTH 2022 SUPPLEMENTAL ICSI, intracytoplasmic sperm injection; IO, induction of ovulation; IQ, intelligence quotient; IUI, intrauterine insemination; IVF, in vitro fertilization; MABC-2, Movement Assessment Battery for Children-Second Edition (Henderson, 2007); MAR, medically assisted reproduction; SD, standard deviation.a MAR corresponds to the whole range of MAR techniques (ie IO, IUI, IVF, and IVF-ICSI); b Chi-squared test P value, estimated with generalized estimating equations approach to consider the correlation between twins and triplets, compared to spontaneous pregnancy; c GYNECOLOGY Original Research

TABLE 6
Association between mode of conception and neurodevelopmental outcome measures at 5.5 years (multivariate complete case analysis)Verhaeghe.Neurodevelopment at age 5 for preterm children born following medically-assisted reproduction.Am J Obstet Gynecol 2022.(continued)Association between mode of conception and neurodevelopmental outcome measures at 5.5 years (multivariate complete case analysis) (continued) GA, gestational age; ICSI, intracytoplasmic sperm injection; IQ, intelligence quotient; IUI, intrauterine insemination; IVF, in vitro fertilization; MABC-2, Movement Assessment Battery for Children-Second Edition (Henderson, 2007); MAR, medically assisted reproduction; MD, mean difference; OI, ovulation induction; OR, odds ratio; SD, standard deviation.MAR corresponds to the whole range of MAR techniques (ie IO, IUI, IVF, and IVF-ICSI); b The reported measures of association are OR, except for Full Scale IQ and total MABC-2 score where MDs are reported.The generalized estimating equations approach is used to consider the correlation between twins or triplets; c Sociodemographic factors adjusted for were maternal age, parity, education level, employment status, living with a partner, smoking during pregnancy, country of birth, and parents' socioeconomic status; d Full Scale IQ, measured by the Wechsler Preschool and Primary Scale of Intelligence-Fourth edition 26 ; e Cutoff of the distribution related to a reference group born at term 27 ; f Severe or moderate cerebral palsy, severe or moderate sensory disabilities, or Full scale IQ <2 SDs below the mean of a reference population; g Among children without cerebral palsy, without severe or moderate sensory disabilities, and with Full Scale IQ !2 SDs below the mean of a reference population.Verhaeghe.Neurodevelopment at age 5 for preterm children born following medically-assisted reproduction.Am J Obstet Gynecol 2022.MONTH 2022 American Journal of Obstetrics & Gynecology 1.e21 CI, confidence interval; a GYNECOLOGY Original Research